Splicing, Folding, and Stretching Nucleic Acids

Williams College, Williamstown MA

Investigators

Linked publications & trials

Abstract

[unreadable] DESCRIPTION (provided by applicant): Single-stranded nucleic acids fold into complex and compact shapes. Making accurate secondary structure predictions helps uncover function and cellular activity and is also of general interest as an analog of the protein folding problem. The following three proposals will expand the predictive power of physical chemical models of nucleic acid conformations. (1) Gene finding algorithms are currently statistics based. We, like cells, favor a thermodynamics-based approach to locating splice sites. Preliminary results show that snRNA binds better to real splice sequences than false. This work should lead to better gene recognition algorithms and a simple picture of splicing. (2) Revising current models and improving the functionality of secondary structure algorithms will help researchers interested in nucleic acid conformations. Algorithmic advances will likely be useful for other related problems such as anti-sense gene therapy and gene chip thermodynamic analysis. (a) Pseudoknots are excluded from popular dynamic-programming algorithms like MFOLD. Characterizing the "PseudoBase" database of known pseudoknots will help select an algorithm, which would then be implemented. (b) Current algorithms search for minimum energy conformations, but secondary structures may be trapped kinetically. A kinetic rule for folding based in polymer physics and implemented in a dynamic programming algorithm would allow the study of misfolding. (c) Unpaired regions are ignored in current models of folding, even though single-strand stacking enthalpies are nearly equal to those of duplex formation. By including these interactions, the accuracy of folding algorithms should improve. (d) Folding algorithms seek the lowest free energy state. Since the free energy differences between several conformations is small, the intrinsic uncertainty of the energy prediction may mean another state is lowest in energy. Characterizing the uncertainty will permit calculations of the likelihood of a correct prediction as a function of the number of nucleotides. (3) The development of our "Stacked or Freely Jointed Chain" (SFJC) model will provide a new context with which to study polymers whose conformational disorder arises from sharp kinks. Calculations are straightforward for this two-state model and preliminary results suggest they compare well with recent stretched DNA experiments. The SFJC is also handy for modeling single-stranded loop sections of RNA. [unreadable] [unreadable]